Days after Anthropic abruptly suspended entry to its newest AI fashions following a US authorities directive, questions proceed to swirl round each the circumstances that led to the Trump administration’s determination and its broader implications for the worldwide AI ecosystem.
Anthropic in a weblog publish on Friday, June 12, mentioned it acquired an export management directive, citing nationwide safety authorities, from the US Commerce Division to cease any international nationwide, whether or not inside or outdoors america, together with its personal international nationwide staff, from accessing Fable 5 and Mythos 5 — two AI fashions underpinned by Mythos Preview, which the AI startup has described as its strongest mannequin but.
To adjust to the directive, Anthropic mentioned that it needed to disable entry to Fable 5 and Mythos 5 for all its clients, together with US residents. The reactions poured in nearly instantly. Throughout Europe and the UK, many seen the transfer as a stark reminder of how closely they depend on US-based AI mannequin suppliers.
Equally, in India, it reignited a long-running debate over technological sovereignty. A number of Indian founders, traders, and coverage specialists took to social media, debating whether or not the nation ought to speed up efforts to construct its personal AI infrastructure, make investments extra aggressively in open-source alternate options, or proceed counting on international frontier fashions. Right here’s a fast roundup of essentially the most notable reactions to date.
Sridhar Vembu, Zoho
Sridhar Vembu, founding father of Indian SaaS firm Zoho, mentioned that the episode exhibits how globalisation is lifeless and that India should discover her personal approach forward. “What can our authorities do proper now? Make sure that orgs in India embrace smaller fashions, each Indian and Chinese language open supply ones,” Vembu wrote on X.
That is large: all entry to Mythos and Fable AI fashions disabled for everybody outdoors America.
First ideas:
1. Expertise is the last word weapon. Nationwide sovereignty, nationwide safety, all of it’s now about know-how.
2. Globalization is lifeless and Bharat should discover her… https://t.co/kCQpq93D3r
— Sridhar Vembu (@svembu) June 13, 2026
“We should deepen our R&D. Sarvam has been on it and we’ve got been on it however do not forget that the most recent fashions value not solely big GPU budgets to coach, the GPUs themselves are restricted […[ I would not like to ask the government to fund tens of billions of dollars on this anyway – the money has far better uses. Zoho has been pursuing alternative R&D approaches that are far, far less expensive but by its nature cutting edge R&D takes time and we are patient. I am confident we will get there,” he further said.
Pratyush Kumar, Sarvam AI
Pratyush Kumar, co-founder and CEO of Sarvam AI, one of the first Indian startups to launch a foundational AI model built from scratch, said, “We need to have more countries and companies owning their own destinies. And in the post AI world, that means being able to use and improve AI systems within their own perimeters – what one may call Sovereign AI.”
Ok, so here is my take on the Fable ban, sovereign AI, Sarvam, etc.
The event is interesting as it has implications from many perspectives.
For AI users, it is clear that you should not confuse access with ownership, or adoption itself as advantage. And if the most significant…
— Pratyush Kumar (@pratykumar) June 13, 2026
“From our vantage point, it is super clear that India will build, leverage, and create massive business value and societal impact with sovereign AI [..] The Fable ban is an efficient instigation for extra individuals to interact in recognising the necessity for sovereignty,” he added.
Aakrit Vaish, Activate
Aakrit Vaish, founding father of Indian AI enterprise platform Activate, opined that Anthropic’s determination “fully modifications issues.” “I feel this materially modifications the best way all of us ought to be fascinated about sovereign AI in India,” he was quoted as saying by JHB.
Story continues under this advert
“Over the previous couple of years, I’d usually argue for sovereign AI by saying “what if at some point the US decides to show the swap off?” Had by no means imagined that day would come so quickly. Sovereign AI has moved from a story to the biggest enterprise alternative of our occasions,” he wrote in a publish on X.
That is completely loopy.
Over the previous couple of years, I’d usually argue for sovereign AI by saying “what if at some point the US decides to show the swap off?”
Had by no means imagined that day would come so quickly.
Sovereign AI has moved from a story to the biggest enterprise… https://t.co/iyjJbjWE3P
— Aakrit Vaish (@aakrit) June 13, 2026
Vaish additionally reportedly plans to encourage firms in his portfolio to scale back their dependence on a small variety of frontier AI suppliers as startups more and more flip to open AI fashions.
Mohandas Pai, investor
Responding to Vembu’s publish on X, investor and former Infosys govt Mohandas Pai argued that the event highlighted the necessity for a much more formidable nationwide AI technique and referred to as on the federal government to considerably enhance investments in AI, computing infrastructure, and deep tech.
“We’re approach behind and wish a nationwide mission to get going shortly. Present govt applications are too gradual, approach too small to make any giant affect. We’d like an annual 50000 cr fund for deep tech and AI, a 200,000 cr ELGS Assure Fund to construct Hyper cloud, {hardware} and chips,” he wrote on X.
For context, the Indian authorities authorised the IndiaAI Mission in 2024 with an outlay of Rs 10, 371 crore over 5 years, aimed toward increasing compute infrastructure, supporting startups, and creating indigenous AI capabilities.
Story continues under this advert
Hemant Mohapatra, Lightspeed
Hemant Mohapatra, a accomplice at VC agency Lightspeed, mentioned, “The “sovereign AI is actual” second is right here. Nation-states will quickly begin needing citizenship and/or safety clearances to work on the subsequent SOTA fashions the best way they do for protection, house, nuclear tech. It’s only a matter of time. Expertise wars right here shall be loopy.”
In response to Pai’s argument that the dearth of capital was India’s main problem in AI, Mohapatra mentioned that the most important constraints to constructing globally aggressive AI firms are expertise, entry to computing sources, and execution.
To coach a GPT class 1T mannequin from scratch – together with failed runs, knowledge acq+clear+rlhf, post-training, crew/individuals will seemingly req $250M of compute on an aggressive 3-4mo schedule (i.e. extra reserved GPUs), $500-600M all-in IF you do a dense one. MoE + fp8 will minimize prices by… https://t.co/KZaFo6BTW8
— Hemant Mohapatra (@MohapatraHemant) June 13, 2026
“To coach a GPT class 1T mannequin from scratch – together with failed runs, knowledge acq+clear+rlhf, post-training, crew/individuals will seemingly req $250M of compute on an aggressive 3-4mo schedule (i.e. extra reserved GPUs), $500-600M all-in IF you do a dense one […] So this isn’t as a lot about elevating $50-60B, or elevating it unexpectedly because the OP says – we’re traders in mistral, sarvam, reflection and anthropic – they usually all scaled capital over time as fashions bought adoption, however the early bottleneck is extra on expertise + GPUs at that scale the place you are able to do attention-grabbing issues,” he wrote in a publish on X.
Samir Saran, ORF
Commenting on evolving export controls, Samir Saran, president of assume tank Observer Analysis Basis (ORF), mentioned, “[Export controls] simply modified form. Not chips or tech, not code. It’s Folks. Overseas nationals, together with a US lab’s personal staff, locked out of frontier fashions in a single day, wherever they sit.”
#Thread#ExportControl simply modified form. Not chips or tech, not code. It’s Folks. Overseas nationals, together with a US lab’s personal staff, locked out of frontier fashions in a single day, wherever they sit.
Hmm… If expertise itself is the leak, ring fencing tech will get loads tougher.
1/n https://t.co/GBOO6Wl7o2— Samir Saran (@samirsaran) June 13, 2026
“[Data sovereignty] is now a sharper negotiation. If entry could be pulled on nationwide safety grounds, the commerce of knowledge will want a tough revision as properly. Leverage sits with whoever holds the swap. Don’t give it away low cost. Concentrate on functionality, not entry,” he added.